imdb/imdb.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "daqPUxKXJgXZ",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "daqPUxKXJgXZ",
    "outputId": "335a6bac-76fd-4d7c-fd73-f80eb34d3151"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: konlpy in d:\\s\\lib\\site-packages (0.6.0)\n",
      "Requirement already satisfied: numpy>=1.6 in d:\\s\\lib\\site-packages (from konlpy) (1.20.3)\n",
      "Requirement already satisfied: lxml>=4.1.0 in d:\\s\\lib\\site-packages (from konlpy) (4.6.3)\n",
      "Requirement already satisfied: JPype1>=0.7.0 in d:\\s\\lib\\site-packages (from konlpy) (1.3.0)\n",
      "Note: you may need to restart the kernel to use updated packages.\n"
     ]
    }
   ],
   "source": [
    "%pip install konlpy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "0111c21b",
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.datasets import imdb\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import re\n",
    "import urllib.request\n",
    "from konlpy.tag import Okt\n",
    "from konlpy.tag import Mecab\n",
    "from tqdm import tqdm\n",
    "from tensorflow.keras.preprocessing.text import Tokenizer\n",
    "from tensorflow.keras.preprocessing.sequence import pad_sequences\n",
    "\n",
    "# 모델 생성\n",
    "import tensorflow.keras as keras\n",
    "from tensorflow.keras.layers import Embedding, Dense, LSTM, Input\n",
    "import tensorflow.keras.models as models\n",
    "from tensorflow.keras.models import Sequential, Model\n",
    "from tensorflow.keras.models import load_model\n",
    "from tensorflow.keras.callbacks import EarlyStopping, ModelCheckpoint\n",
    "import tensorflow.keras.layers as layers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "4704a750",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "4704a750",
    "outputId": "4b56a6d9-c91c-4909-8685-31dd8d2d04ef"
   },
   "outputs": [],
   "source": [
    "(X_train, Y_train), (X_test, Y_test) = imdb.load_data(num_words=20000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "523fbfc0",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "523fbfc0",
    "outputId": "4ea816de-425a-4faf-ef1a-a496782475d3"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "훈련 데이터: 25000\n",
      "검증(테스트) 데이터: 25000\n",
      "카테고리: 2\n"
     ]
    }
   ],
   "source": [
    "print(f\"훈련 데이터: {len(X_train)}\")\n",
    "print(f\"검증(테스트) 데이터: {len(X_test)}\")\n",
    "num_classes = len(set(Y_train))\n",
    "print(f\"카테고리: {num_classes}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "8268fb59",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "8268fb59",
    "outputId": "f53fbb04-6652-4600-cf4f-67ebe90ac2ad"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[1,\n",
       " 14,\n",
       " 22,\n",
       " 16,\n",
       " 43,\n",
       " 530,\n",
       " 973,\n",
       " 1622,\n",
       " 1385,\n",
       " 65,\n",
       " 458,\n",
       " 4468,\n",
       " 66,\n",
       " 3941,\n",
       " 4,\n",
       " 173,\n",
       " 36,\n",
       " 256,\n",
       " 5,\n",
       " 25,\n",
       " 100,\n",
       " 43,\n",
       " 838,\n",
       " 112,\n",
       " 50,\n",
       " 670,\n",
       " 2,\n",
       " 9,\n",
       " 35,\n",
       " 480,\n",
       " 284,\n",
       " 5,\n",
       " 150,\n",
       " 4,\n",
       " 172,\n",
       " 112,\n",
       " 167,\n",
       " 2,\n",
       " 336,\n",
       " 385,\n",
       " 39,\n",
       " 4,\n",
       " 172,\n",
       " 4536,\n",
       " 1111,\n",
       " 17,\n",
       " 546,\n",
       " 38,\n",
       " 13,\n",
       " 447,\n",
       " 4,\n",
       " 192,\n",
       " 50,\n",
       " 16,\n",
       " 6,\n",
       " 147,\n",
       " 2025,\n",
       " 19,\n",
       " 14,\n",
       " 22,\n",
       " 4,\n",
       " 1920,\n",
       " 4613,\n",
       " 469,\n",
       " 4,\n",
       " 22,\n",
       " 71,\n",
       " 87,\n",
       " 12,\n",
       " 16,\n",
       " 43,\n",
       " 530,\n",
       " 38,\n",
       " 76,\n",
       " 15,\n",
       " 13,\n",
       " 1247,\n",
       " 4,\n",
       " 22,\n",
       " 17,\n",
       " 515,\n",
       " 17,\n",
       " 12,\n",
       " 16,\n",
       " 626,\n",
       " 18,\n",
       " 19193,\n",
       " 5,\n",
       " 62,\n",
       " 386,\n",
       " 12,\n",
       " 8,\n",
       " 316,\n",
       " 8,\n",
       " 106,\n",
       " 5,\n",
       " 4,\n",
       " 2223,\n",
       " 5244,\n",
       " 16,\n",
       " 480,\n",
       " 66,\n",
       " 3785,\n",
       " 33,\n",
       " 4,\n",
       " 130,\n",
       " 12,\n",
       " 16,\n",
       " 38,\n",
       " 619,\n",
       " 5,\n",
       " 25,\n",
       " 124,\n",
       " 51,\n",
       " 36,\n",
       " 135,\n",
       " 48,\n",
       " 25,\n",
       " 1415,\n",
       " 33,\n",
       " 6,\n",
       " 22,\n",
       " 12,\n",
       " 215,\n",
       " 28,\n",
       " 77,\n",
       " 52,\n",
       " 5,\n",
       " 14,\n",
       " 407,\n",
       " 16,\n",
       " 82,\n",
       " 10311,\n",
       " 8,\n",
       " 4,\n",
       " 107,\n",
       " 117,\n",
       " 5952,\n",
       " 15,\n",
       " 256,\n",
       " 4,\n",
       " 2,\n",
       " 7,\n",
       " 3766,\n",
       " 5,\n",
       " 723,\n",
       " 36,\n",
       " 71,\n",
       " 43,\n",
       " 530,\n",
       " 476,\n",
       " 26,\n",
       " 400,\n",
       " 317,\n",
       " 46,\n",
       " 7,\n",
       " 4,\n",
       " 12118,\n",
       " 1029,\n",
       " 13,\n",
       " 104,\n",
       " 88,\n",
       " 4,\n",
       " 381,\n",
       " 15,\n",
       " 297,\n",
       " 98,\n",
       " 32,\n",
       " 2071,\n",
       " 56,\n",
       " 26,\n",
       " 141,\n",
       " 6,\n",
       " 194,\n",
       " 7486,\n",
       " 18,\n",
       " 4,\n",
       " 226,\n",
       " 22,\n",
       " 21,\n",
       " 134,\n",
       " 476,\n",
       " 26,\n",
       " 480,\n",
       " 5,\n",
       " 144,\n",
       " 30,\n",
       " 5535,\n",
       " 18,\n",
       " 51,\n",
       " 36,\n",
       " 28,\n",
       " 224,\n",
       " 92,\n",
       " 25,\n",
       " 104,\n",
       " 4,\n",
       " 226,\n",
       " 65,\n",
       " 16,\n",
       " 38,\n",
       " 1334,\n",
       " 88,\n",
       " 12,\n",
       " 16,\n",
       " 283,\n",
       " 5,\n",
       " 16,\n",
       " 4472,\n",
       " 113,\n",
       " 103,\n",
       " 32,\n",
       " 15,\n",
       " 16,\n",
       " 5345,\n",
       " 19,\n",
       " 178,\n",
       " 32]"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "5f66af6a",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "5f66af6a",
    "outputId": "8a076db7-8e65-46f0-b0b4-7dc866f72a69"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "19999"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "max([max(sequence) for sequence in X_train])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "68f2a372",
   "metadata": {
    "id": "68f2a372"
   },
   "outputs": [],
   "source": [
    "from keras.models import load_model\n",
    "from keras.layers import Lambda\n",
    "import tensorflow as tf"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "8d533ea1",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "8d533ea1",
    "outputId": "f1d59424-6eae-4eb1-d79f-a1231b3d3feb"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "the as you with out themselves powerful lets loves their becomes reaching had journalist of lot from anyone to have after out atmosphere never more room and it so heart shows to years of every never going and help moments or of every chest visual movie except her was several of enough more with is now current film as you of mine potentially unfortunately of you than him that with out themselves her get for was camp of you movie sometimes movie that with scary but pratfalls to story wonderful that in seeing in character to of 70s musicians with heart had shadows they of here that with her serious to have does when from why what have critics they is you that isn't one will very to as itself with other tricky in of seen over landed for anyone of and br show's to whether from than out themselves history he name half some br of 'n odd was two most of mean for 1 any an boat she he should is thought frog but of script you not while history he heart to real at barrel but when from one bit then have two of script their with her nobody most that with wasn't to with armed acting watch an for with heartfelt film want an\n"
     ]
    }
   ],
   "source": [
    "word_to_idx = tf.keras.datasets.imdb.get_word_index()\n",
    "idx_to_word = dict([(value, key) for (key, value) in word_to_idx.items()])\n",
    "decoded_review = ' '.join([idx_to_word.get(i) for i in X_train[0]])\n",
    "print(decoded_review)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "5328e958",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "5328e958",
    "outputId": "57d013c6-84b8-476e-aa56-ccb9b82b9996"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(25000, 20000)\n",
      "(25000, 20000)\n"
     ]
    }
   ],
   "source": [
    "def vectorize_sequences(sequences, dimension=20000):\n",
    "    # Create an all-zero matrix of shape (len(sequences), dimension)\n",
    "    results = np.zeros((len(sequences), dimension))\n",
    "    for i, sequence in enumerate(sequences):\n",
    "       results[i, sequence] = 1.  # set specific indices of results[i] to 1s\n",
    "    return results\n",
    "X_train = vectorize_sequences(X_train)\n",
    "X_test = vectorize_sequences(X_test)\n",
    "print(X_train.shape)\n",
    "print(X_test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "681bb6ec",
   "metadata": {
    "id": "681bb6ec"
   },
   "outputs": [],
   "source": [
    "Y_train = np.array(Y_train).astype('float32')\n",
    "Y_test = np.array(Y_test).astype('float32')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "5cfb31c0",
   "metadata": {
    "id": "5cfb31c0"
   },
   "outputs": [],
   "source": [
    "max_features = 20000 # 가장 빈번하게 사용된 5000개의 단어만 사용\n",
    "maxlen = 500 # 각 시퀀스의 최대 길이\n",
    "(X_train, Y_train), (X_test, Y_test) = keras.datasets.imdb.load_data(num_words=max_features)\n",
    "# 시퀀스의 길이를 maxlen으로 변경해줌(maxlen보다 짧았으면, 0으로 채워준다)\n",
    "X_train = keras.preprocessing.sequence.pad_sequences(X_train, maxlen=maxlen, padding='post')\n",
    "X_test = keras.preprocessing.sequence.pad_sequences(X_test, maxlen=maxlen, padding='post')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f30aa797",
   "metadata": {
    "id": "f30aa797"
   },
   "source": [
    "## Word Embedding layer를 이용한 Dense model1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "2d7c9976",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "2d7c9976",
    "outputId": "0266a651-bb98-4560-f8fa-32be64d6b4d4"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " embedding (Embedding)       (None, 500, 64)           1280000   \n",
      "                                                                 \n",
      " flatten (Flatten)           (None, 32000)             0         \n",
      "                                                                 \n",
      " dense (Dense)               (None, 1)                 32001     \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 1,312,001\n",
      "Trainable params: 1,312,001\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model1 = models.Sequential(\n",
    "[layers.Embedding(20000, 64, input_length=maxlen),\n",
    "layers.Flatten(),\n",
    "layers.Dense(1, activation='sigmoid')])\n",
    "model1.compile(optimizer='adam', loss='binary_crossentropy', metrics='acc')\n",
    "model1.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "f2d7ea26",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "f2d7ea26",
    "outputId": "2f249d61-c169-4dbb-88f4-34a2198c43e0"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "274/274 [==============================] - 4s 15ms/step - loss: 0.5444 - acc: 0.7162 - val_loss: 0.3399 - val_acc: 0.8616\n",
      "Epoch 2/10\n",
      "274/274 [==============================] - 4s 15ms/step - loss: 0.2175 - acc: 0.9225 - val_loss: 0.2797 - val_acc: 0.8860\n",
      "Epoch 3/10\n",
      "274/274 [==============================] - 4s 15ms/step - loss: 0.0953 - acc: 0.9770 - val_loss: 0.2813 - val_acc: 0.8896\n",
      "Epoch 4/10\n",
      "274/274 [==============================] - 4s 15ms/step - loss: 0.0405 - acc: 0.9950 - val_loss: 0.2920 - val_acc: 0.8888\n",
      "Epoch 5/10\n",
      "274/274 [==============================] - 4s 15ms/step - loss: 0.0190 - acc: 0.9990 - val_loss: 0.3104 - val_acc: 0.8852\n",
      "Epoch 6/10\n",
      "274/274 [==============================] - 4s 15ms/step - loss: 0.0106 - acc: 0.9997 - val_loss: 0.3230 - val_acc: 0.8885\n",
      "Epoch 7/10\n",
      "274/274 [==============================] - 4s 16ms/step - loss: 0.0064 - acc: 0.9999 - val_loss: 0.3366 - val_acc: 0.8869\n",
      "Epoch 8/10\n",
      "274/274 [==============================] - 4s 15ms/step - loss: 0.0042 - acc: 0.9999 - val_loss: 0.3486 - val_acc: 0.8859\n",
      "Epoch 9/10\n",
      "274/274 [==============================] - 4s 15ms/step - loss: 0.0029 - acc: 1.0000 - val_loss: 0.3616 - val_acc: 0.8853\n",
      "Epoch 10/10\n",
      "274/274 [==============================] - 4s 15ms/step - loss: 0.0021 - acc: 1.0000 - val_loss: 0.3705 - val_acc: 0.8860\n"
     ]
    }
   ],
   "source": [
    "history = model1.fit(X_train, Y_train,\n",
    "epochs=10,\n",
    "batch_size=64,\n",
    "validation_split=0.3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "69273ac0",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 111
    },
    "id": "69273ac0",
    "outputId": "4406a4d9-b197-4be1-8cb9-5fb2a3058753"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'\\nEpoch 1/10\\n157/157 [==============================] - 2s 14ms/step - loss: 0.6308 - acc: 0.6414 - val_loss: 0.4587 - val_acc: 0.8152\\nEpoch 2/10\\n157/157 [==============================] - 2s 14ms/step - loss: 0.3258 - acc: 0.8787 - val_loss: 0.3098 - val_acc: 0.8758\\nEpoch 3/10\\n157/157 [==============================] - 2s 14ms/step - loss: 0.2055 - acc: 0.9290 - val_loss: 0.2883 - val_acc: 0.8830\\nEpoch 4/10\\n157/157 [==============================] - 2s 14ms/step - loss: 0.1417 - acc: 0.9578 - val_loss: 0.2924 - val_acc: 0.8846\\nEpoch 5/10\\n157/157 [==============================] - 2s 13ms/step - loss: 0.0959 - acc: 0.9778 - val_loss: 0.2965 - val_acc: 0.8814\\nEpoch 6/10\\n157/157 [==============================] - 2s 13ms/step - loss: 0.0631 - acc: 0.9900 - val_loss: 0.3086 - val_acc: 0.8804\\nEpoch 7/10\\n157/157 [==============================] - 2s 12ms/step - loss: 0.0422 - acc: 0.9959 - val_loss: 0.3223 - val_acc: 0.8768\\nEpoch 8/10\\n157/157 [==============================] - 2s 13ms/step - loss: 0.0287 - acc: 0.9983 - val_loss: 0.3356 - val_acc: 0.8758\\nEpoch 9/10\\n157/157 [==============================] - 2s 14ms/step - loss: 0.0202 - acc: 0.9992 - val_loss: 0.3497 - val_acc: 0.8766\\nEpoch 10/10\\n157/157 [==============================] - 2s 13ms/step - loss: 0.0149 - acc: 0.9995 - val_loss: 0.3619 - val_acc: 0.8754\\n'"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#sigmoid 10000,32, batch_size=128\n",
    "'''\n",
    "Epoch 1/10\n",
    "157/157 [==============================] - 2s 14ms/step - loss: 0.6308 - acc: 0.6414 - val_loss: 0.4587 - val_acc: 0.8152\n",
    "Epoch 2/10\n",
    "157/157 [==============================] - 2s 14ms/step - loss: 0.3258 - acc: 0.8787 - val_loss: 0.3098 - val_acc: 0.8758\n",
    "Epoch 3/10\n",
    "157/157 [==============================] - 2s 14ms/step - loss: 0.2055 - acc: 0.9290 - val_loss: 0.2883 - val_acc: 0.8830\n",
    "Epoch 4/10\n",
    "157/157 [==============================] - 2s 14ms/step - loss: 0.1417 - acc: 0.9578 - val_loss: 0.2924 - val_acc: 0.8846\n",
    "Epoch 5/10\n",
    "157/157 [==============================] - 2s 13ms/step - loss: 0.0959 - acc: 0.9778 - val_loss: 0.2965 - val_acc: 0.8814\n",
    "Epoch 6/10\n",
    "157/157 [==============================] - 2s 13ms/step - loss: 0.0631 - acc: 0.9900 - val_loss: 0.3086 - val_acc: 0.8804\n",
    "Epoch 7/10\n",
    "157/157 [==============================] - 2s 12ms/step - loss: 0.0422 - acc: 0.9959 - val_loss: 0.3223 - val_acc: 0.8768\n",
    "Epoch 8/10\n",
    "157/157 [==============================] - 2s 13ms/step - loss: 0.0287 - acc: 0.9983 - val_loss: 0.3356 - val_acc: 0.8758\n",
    "Epoch 9/10\n",
    "157/157 [==============================] - 2s 14ms/step - loss: 0.0202 - acc: 0.9992 - val_loss: 0.3497 - val_acc: 0.8766\n",
    "Epoch 10/10\n",
    "157/157 [==============================] - 2s 13ms/step - loss: 0.0149 - acc: 0.9995 - val_loss: 0.3619 - val_acc: 0.8754\n",
    "'''"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "c9decc79",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 541
    },
    "id": "c9decc79",
    "outputId": "39c3500f-82e2-46c4-b40f-5564b4240ead"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#그래프 \n",
    "import matplotlib.pyplot as plt\n",
    "plt.plot(history.history['loss'], 'b-', label='train loss')\n",
    "plt.plot(history.history['val_loss'], 'b--', label='valid loss')\n",
    "plt.xlabel('Epochs')\n",
    "plt.ylabel('Loss')\n",
    "plt.legend()\n",
    "plt.figure()\n",
    "\n",
    "\n",
    "plt.plot(history.history['acc'], 'g-', label='train acc')\n",
    "plt.plot(history.history['val_acc'], 'g--', label='valid acc')\n",
    "plt.xlabel('Epochs')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "260bbf62",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "260bbf62",
    "outputId": "d3c3f410-99d3-4556-d32a-985d7e32cb04"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "782/782 [==============================] - 1s 1ms/step - loss: 0.4150 - acc: 0.8651\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[0.4149598777294159, 0.8650799989700317]"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model1.evaluate(X_test, Y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9kSlWZGgMYM5",
   "metadata": {
    "id": "9kSlWZGgMYM5"
   },
   "source": [
    "### SimpleRNN layer model2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "T7aDJDUMQs0g",
   "metadata": {
    "id": "T7aDJDUMQs0g"
   },
   "outputs": [],
   "source": [
    "max_features = 5000 # 가장 빈번하게 사용된 5000개의 단어만 사용\n",
    "maxlen = 500 # 각 시퀀스의 최대 길이\n",
    "(X_train, Y_train), (X_test, Y_test) = keras.datasets.imdb.load_data(num_words=max_features)\n",
    "# 시퀀스의 길이를 maxlen으로 변경해줌(maxlen보다 짧았으면, 0으로 채워준다)\n",
    "X_train = keras.preprocessing.sequence.pad_sequences(X_train, maxlen=maxlen, padding='post')\n",
    "X_test = keras.preprocessing.sequence.pad_sequences(X_test, maxlen=maxlen, padding='post')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "OVvqRKXFNUTG",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "OVvqRKXFNUTG",
    "outputId": "1a7ccde1-c72b-4ddc-a871-3e0ea441c6c9"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_1\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " embedding_1 (Embedding)     (None, None, 1000)        5000000   \n",
      "                                                                 \n",
      " simple_rnn (SimpleRNN)      (None, 64)                68160     \n",
      "                                                                 \n",
      " dense_1 (Dense)             (None, 32)                2080      \n",
      "                                                                 \n",
      " dropout_1 (Dropout)         (None, 32)                0         \n",
      "                                                                 \n",
      " dense_2 (Dense)             (None, 1)                 33        \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 5,070,273\n",
      "Trainable params: 5,070,273\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model2 = models.Sequential(\n",
    "[layers.Embedding(max_features, 1000),\n",
    "layers.SimpleRNN(64),\n",
    "layers.Dense(32, activation='softmax'),\n",
    "layers.Dropout(0.2),\n",
    "layers.Dense(1, activation='sigmoid')]\n",
    ")\n",
    "model2.compile(optimizer='adam', loss='binary_crossentropy', metrics='acc')\n",
    "model2.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "64e685d1",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "313/313 [==============================] - 5s 15ms/step - loss: 0.0859 - acc: 0.9737 - val_loss: 0.3980 - val_acc: 0.8682\n",
      "Epoch 2/10\n",
      "313/313 [==============================] - 5s 14ms/step - loss: 0.0131 - acc: 0.9993 - val_loss: 0.4121 - val_acc: 0.8694\n",
      "Epoch 3/10\n",
      "313/313 [==============================] - 5s 15ms/step - loss: 0.0062 - acc: 0.9999 - val_loss: 0.4282 - val_acc: 0.8686\n",
      "Epoch 4/10\n",
      "313/313 [==============================] - 5s 14ms/step - loss: 0.0042 - acc: 0.9999 - val_loss: 0.4423 - val_acc: 0.8702\n",
      "Epoch 5/10\n",
      "313/313 [==============================] - 5s 14ms/step - loss: 0.0030 - acc: 1.0000 - val_loss: 0.4561 - val_acc: 0.8698\n",
      "Epoch 6/10\n",
      "313/313 [==============================] - 5s 15ms/step - loss: 0.0023 - acc: 1.0000 - val_loss: 0.4693 - val_acc: 0.8692\n",
      "Epoch 7/10\n",
      "313/313 [==============================] - 5s 15ms/step - loss: 0.0018 - acc: 1.0000 - val_loss: 0.4817 - val_acc: 0.8692\n",
      "Epoch 8/10\n",
      "313/313 [==============================] - 5s 15ms/step - loss: 0.0014 - acc: 1.0000 - val_loss: 0.4941 - val_acc: 0.8690\n",
      "Epoch 9/10\n",
      "313/313 [==============================] - 5s 15ms/step - loss: 0.0011 - acc: 1.0000 - val_loss: 0.5040 - val_acc: 0.8690\n",
      "Epoch 10/10\n",
      "313/313 [==============================] - 5s 15ms/step - loss: 8.8533e-04 - acc: 1.0000 - val_loss: 0.5161 - val_acc: 0.8694\n"
     ]
    }
   ],
   "source": [
    "r = model1.fit(X_train, Y_train,\n",
    "epochs=10,\n",
    "batch_size=64,\n",
    "validation_split=0.2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "-Nn8THF1PiZf",
   "metadata": {
    "id": "-Nn8THF1PiZf"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'\\nEpoch 1/10\\n274/274 [==============================] - 8s 28ms/step - loss: 0.0016 - acc: 1.0000 - val_loss: 0.3873 - val_acc: 0.8828\\nEpoch 2/10\\n274/274 [==============================] - 7s 27ms/step - loss: 0.0012 - acc: 1.0000 - val_loss: 0.3939 - val_acc: 0.8840\\nEpoch 3/10\\n274/274 [==============================] - 7s 27ms/step - loss: 9.4417e-04 - acc: 1.0000 - val_loss: 0.4031 - val_acc: 0.8845\\nEpoch 4/10\\n274/274 [==============================] - 8s 28ms/step - loss: 7.5166e-04 - acc: 1.0000 - val_loss: 0.4109 - val_acc: 0.8831\\nEpoch 5/10\\n274/274 [==============================] - 7s 27ms/step - loss: 5.9703e-04 - acc: 1.0000 - val_loss: 0.4191 - val_acc: 0.8835\\nEpoch 6/10\\n274/274 [==============================] - 7s 27ms/step - loss: 4.8367e-04 - acc: 1.0000 - val_loss: 0.4279 - val_acc: 0.8828\\nEpoch 7/10\\n274/274 [==============================] - 8s 27ms/step - loss: 3.9593e-04 - acc: 1.0000 - val_loss: 0.4358 - val_acc: 0.8828\\nEpoch 8/10\\n274/274 [==============================] - 7s 26ms/step - loss: 3.2646e-04 - acc: 1.0000 - val_loss: 0.4420 - val_acc: 0.8840\\nEpoch 9/10\\n274/274 [==============================] - 7s 27ms/step - loss: 2.6959e-04 - acc: 1.0000 - val_loss: 0.4516 - val_acc: 0.8824\\nEpoch 10/10\\n274/274 [==============================] - 7s 27ms/step - loss: 2.2508e-04 - acc: 1.0000 - val_loss: 0.4581 - val_acc: 0.8819\\n'"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#sigmoid 5000,500, batch_size=64\n",
    "'''\n",
    "Epoch 1/10\n",
    "274/274 [==============================] - 8s 28ms/step - loss: 0.0016 - acc: 1.0000 - val_loss: 0.3873 - val_acc: 0.8828\n",
    "Epoch 2/10\n",
    "274/274 [==============================] - 7s 27ms/step - loss: 0.0012 - acc: 1.0000 - val_loss: 0.3939 - val_acc: 0.8840\n",
    "Epoch 3/10\n",
    "274/274 [==============================] - 7s 27ms/step - loss: 9.4417e-04 - acc: 1.0000 - val_loss: 0.4031 - val_acc: 0.8845\n",
    "Epoch 4/10\n",
    "274/274 [==============================] - 8s 28ms/step - loss: 7.5166e-04 - acc: 1.0000 - val_loss: 0.4109 - val_acc: 0.8831\n",
    "Epoch 5/10\n",
    "274/274 [==============================] - 7s 27ms/step - loss: 5.9703e-04 - acc: 1.0000 - val_loss: 0.4191 - val_acc: 0.8835\n",
    "Epoch 6/10\n",
    "274/274 [==============================] - 7s 27ms/step - loss: 4.8367e-04 - acc: 1.0000 - val_loss: 0.4279 - val_acc: 0.8828\n",
    "Epoch 7/10\n",
    "274/274 [==============================] - 8s 27ms/step - loss: 3.9593e-04 - acc: 1.0000 - val_loss: 0.4358 - val_acc: 0.8828\n",
    "Epoch 8/10\n",
    "274/274 [==============================] - 7s 26ms/step - loss: 3.2646e-04 - acc: 1.0000 - val_loss: 0.4420 - val_acc: 0.8840\n",
    "Epoch 9/10\n",
    "274/274 [==============================] - 7s 27ms/step - loss: 2.6959e-04 - acc: 1.0000 - val_loss: 0.4516 - val_acc: 0.8824\n",
    "Epoch 10/10\n",
    "274/274 [==============================] - 7s 27ms/step - loss: 2.2508e-04 - acc: 1.0000 - val_loss: 0.4581 - val_acc: 0.8819\n",
    "'''"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "l0w8OFe0RFJv",
   "metadata": {
    "id": "l0w8OFe0RFJv"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#그래프 \n",
    "import matplotlib.pyplot as plt\n",
    "plt.plot(r.history['loss'], 'b-', label='train loss')\n",
    "plt.plot(r.history['val_loss'], 'b--', label='valid loss')\n",
    "plt.xlabel('Epochs')\n",
    "plt.ylabel('Loss')\n",
    "plt.legend()\n",
    "plt.figure()\n",
    "\n",
    "\n",
    "plt.plot(r.history['acc'], 'g-', label='train acc')\n",
    "plt.plot(r.history['val_acc'], 'g--', label='valid acc')\n",
    "plt.xlabel('Epochs')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "EZSjGc_oRPWz",
   "metadata": {
    "id": "EZSjGc_oRPWz"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "782/782 [==============================] - 53s 68ms/step - loss: 0.6932 - acc: 0.5003\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[0.6931757926940918, 0.5003200173377991]"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model2.evaluate(X_test, Y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ncu7RGmuTfIh",
   "metadata": {
    "id": "ncu7RGmuTfIh"
   },
   "source": [
    "### LSTM layer model3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "0rcJS_lRTjDC",
   "metadata": {
    "id": "0rcJS_lRTjDC"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_4\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " embedding_4 (Embedding)     (None, None, 500)         2500000   \n",
      "                                                                 \n",
      " lstm_2 (LSTM)               (None, 64)                144640    \n",
      "                                                                 \n",
      " dense_5 (Dense)             (None, 1)                 65        \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 2,644,705\n",
      "Trainable params: 2,644,705\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model3 = models.Sequential(\n",
    "[layers.Embedding(max_features, 500),\n",
    "layers.LSTM(64),\n",
    "layers.Dense(1, activation='sigmoid')]\n",
    ")\n",
    "model3.compile(optimizer='adam', loss='binary_crossentropy', metrics='acc')\n",
    "model3.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "LO6JqaiqUbwk",
   "metadata": {
    "id": "LO6JqaiqUbwk"
   },
   "outputs": [],
   "source": [
    "es = EarlyStopping(monitor='val_loss', mode='min', verbose=1, patience=4)\n",
    "mc = ModelCheckpoint(\"review_model.h5\", monitor='val_acc', mode='max', verbose=1, save_best_only=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "egu_HENKd51X",
   "metadata": {
    "id": "egu_HENKd51X"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "625/625 [==============================] - ETA: 0s - loss: 0.6918 - acc: 0.5080\n",
      "Epoch 1: val_acc improved from -inf to 0.51600, saving model to review_model.h5\n",
      "625/625 [==============================] - 132s 210ms/step - loss: 0.6918 - acc: 0.5080 - val_loss: 0.6877 - val_acc: 0.5160\n",
      "Epoch 2/10\n",
      "625/625 [==============================] - ETA: 0s - loss: 0.6744 - acc: 0.5202\n",
      "Epoch 2: val_acc improved from 0.51600 to 0.52040, saving model to review_model.h5\n",
      "625/625 [==============================] - 131s 210ms/step - loss: 0.6744 - acc: 0.5202 - val_loss: 0.6882 - val_acc: 0.5204\n",
      "Epoch 3/10\n",
      "625/625 [==============================] - ETA: 0s - loss: 0.6515 - acc: 0.5367\n",
      "Epoch 3: val_acc did not improve from 0.52040\n",
      "625/625 [==============================] - 132s 212ms/step - loss: 0.6515 - acc: 0.5367 - val_loss: 0.7023 - val_acc: 0.5112\n",
      "Epoch 4/10\n",
      "625/625 [==============================] - ETA: 0s - loss: 0.6398 - acc: 0.5415\n",
      "Epoch 4: val_acc did not improve from 0.52040\n",
      "625/625 [==============================] - 131s 209ms/step - loss: 0.6398 - acc: 0.5415 - val_loss: 0.7134 - val_acc: 0.5182\n",
      "Epoch 5/10\n",
      "625/625 [==============================] - ETA: 0s - loss: 0.4972 - acc: 0.7319\n",
      "Epoch 5: val_acc improved from 0.52040 to 0.85020, saving model to review_model.h5\n",
      "625/625 [==============================] - 132s 212ms/step - loss: 0.4972 - acc: 0.7319 - val_loss: 0.3494 - val_acc: 0.8502\n",
      "Epoch 6/10\n",
      "625/625 [==============================] - ETA: 0s - loss: 0.2554 - acc: 0.8982\n",
      "Epoch 6: val_acc improved from 0.85020 to 0.87840, saving model to review_model.h5\n",
      "625/625 [==============================] - 133s 213ms/step - loss: 0.2554 - acc: 0.8982 - val_loss: 0.3119 - val_acc: 0.8784\n",
      "Epoch 7/10\n",
      "625/625 [==============================] - ETA: 0s - loss: 0.1807 - acc: 0.9311\n",
      "Epoch 7: val_acc improved from 0.87840 to 0.88200, saving model to review_model.h5\n",
      "625/625 [==============================] - 133s 213ms/step - loss: 0.1807 - acc: 0.9311 - val_loss: 0.3103 - val_acc: 0.8820\n",
      "Epoch 8/10\n",
      "625/625 [==============================] - ETA: 0s - loss: 0.1239 - acc: 0.9578\n",
      "Epoch 8: val_acc did not improve from 0.88200\n",
      "625/625 [==============================] - 133s 214ms/step - loss: 0.1239 - acc: 0.9578 - val_loss: 0.3560 - val_acc: 0.8748\n",
      "Epoch 9/10\n",
      "625/625 [==============================] - ETA: 0s - loss: 0.0817 - acc: 0.9744\n",
      "Epoch 9: val_acc did not improve from 0.88200\n",
      "625/625 [==============================] - 131s 210ms/step - loss: 0.0817 - acc: 0.9744 - val_loss: 0.4233 - val_acc: 0.8748\n",
      "Epoch 10/10\n",
      "625/625 [==============================] - ETA: 0s - loss: 0.0570 - acc: 0.9830\n",
      "Epoch 10: val_acc did not improve from 0.88200\n",
      "625/625 [==============================] - 133s 213ms/step - loss: 0.0570 - acc: 0.9830 - val_loss: 0.4564 - val_acc: 0.8726\n"
     ]
    }
   ],
   "source": [
    "r = model3.fit(X_train, Y_train, epochs=10, validation_split=0.2, callbacks=[es, mc])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "05Kk4Jd3hTQ0",
   "metadata": {
    "id": "05Kk4Jd3hTQ0"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#그래프 \n",
    "import matplotlib.pyplot as plt\n",
    "plt.plot(r.history['loss'], 'b-', label='train loss')\n",
    "plt.plot(r.history['val_loss'], 'b--', label='valid loss')\n",
    "plt.xlabel('Epochs')\n",
    "plt.ylabel('Loss')\n",
    "plt.legend()\n",
    "plt.figure()\n",
    "\n",
    "\n",
    "plt.plot(r.history['acc'], 'g-', label='train acc')\n",
    "plt.plot(r.history['val_acc'], 'g--', label='valid acc')\n",
    "plt.xlabel('Epochs')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "ioqVxBDNj90a",
   "metadata": {
    "id": "ioqVxBDNj90a"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "782/782 [==============================] - 50s 63ms/step - loss: 0.4708 - acc: 0.8662\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[0.47076937556266785, 0.8662400245666504]"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model3.evaluate(X_test, Y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "38c6c4bf",
   "metadata": {},
   "source": [
    "### GRU layer model4"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "6867ae67",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_9\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " embedding_9 (Embedding)     (None, None, 100)         500000    \n",
      "                                                                 \n",
      " gru_4 (GRU)                 (None, 128)               88320     \n",
      "                                                                 \n",
      " dense_10 (Dense)            (None, 1)                 129       \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 588,449\n",
      "Trainable params: 588,449\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model4 = models.Sequential(\n",
    "[layers.Embedding(5000, 100),\n",
    "layers.GRU(128),\n",
    "layers.Dense(1, activation='sigmoid')]\n",
    ")\n",
    "model4.compile(optimizer='adam', loss='binary_crossentropy', metrics='acc')\n",
    "model4.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "e2cc821c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "274/274 [==============================] - 106s 382ms/step - loss: 0.6927 - acc: 0.5029 - val_loss: 0.6921 - val_acc: 0.5127\n",
      "Epoch 2/10\n",
      "274/274 [==============================] - 105s 382ms/step - loss: 0.6855 - acc: 0.5114 - val_loss: 0.6901 - val_acc: 0.5013\n",
      "Epoch 3/10\n",
      "274/274 [==============================] - 106s 386ms/step - loss: 0.7547 - acc: 0.5166 - val_loss: 0.6990 - val_acc: 0.4952\n",
      "Epoch 4/10\n",
      "274/274 [==============================] - 106s 388ms/step - loss: 0.6872 - acc: 0.5183 - val_loss: 0.6913 - val_acc: 0.5037\n",
      "Epoch 5/10\n",
      "274/274 [==============================] - 109s 398ms/step - loss: 0.6783 - acc: 0.5468 - val_loss: 0.6830 - val_acc: 0.5592\n",
      "Epoch 6/10\n",
      "274/274 [==============================] - 110s 401ms/step - loss: 0.6568 - acc: 0.6000 - val_loss: 0.6320 - val_acc: 0.6532\n",
      "Epoch 7/10\n",
      "274/274 [==============================] - 110s 400ms/step - loss: 0.5127 - acc: 0.7542 - val_loss: 0.4993 - val_acc: 0.7632\n",
      "Epoch 8/10\n",
      "274/274 [==============================] - 111s 405ms/step - loss: 0.3769 - acc: 0.8411 - val_loss: 0.4006 - val_acc: 0.8284\n",
      "Epoch 9/10\n",
      "274/274 [==============================] - 110s 403ms/step - loss: 0.2677 - acc: 0.8963 - val_loss: 0.4080 - val_acc: 0.8257\n",
      "Epoch 10/10\n",
      "274/274 [==============================] - 111s 404ms/step - loss: 0.2076 - acc: 0.9261 - val_loss: 0.3478 - val_acc: 0.8653\n"
     ]
    }
   ],
   "source": [
    "r = model4.fit(X_train, Y_train,\n",
    "epochs=10,\n",
    "batch_size=64,\n",
    "validation_split=0.3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "0860e4a8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#그래프 \n",
    "import matplotlib.pyplot as plt\n",
    "plt.plot(r.history['loss'], 'b-', label='train loss')\n",
    "plt.plot(r.history['val_loss'], 'b--', label='valid loss')\n",
    "plt.xlabel('Epochs')\n",
    "plt.ylabel('Loss')\n",
    "plt.legend()\n",
    "plt.figure()\n",
    "\n",
    "\n",
    "plt.plot(r.history['acc'], 'g-', label='train acc')\n",
    "plt.plot(r.history['val_acc'], 'g--', label='valid acc')\n",
    "plt.xlabel('Epochs')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "b51ad103",
   "metadata": {},
   "outputs": [],
   "source": [
    "es = EarlyStopping(monitor='val_loss', mode='min', verbose=1, patience=4)\n",
    "mc = ModelCheckpoint(\"review_model.h5\", monitor='val_acc', mode='max', verbose=1, save_best_only=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "60765c99",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "625/625 [==============================] - ETA: 0s - loss: 0.2071 - acc: 0.9222\n",
      "Epoch 1: val_acc improved from -inf to 0.87740, saving model to review_model.h5\n",
      "625/625 [==============================] - 113s 179ms/step - loss: 0.2071 - acc: 0.9222 - val_loss: 0.3122 - val_acc: 0.8774\n",
      "Epoch 2/10\n",
      "625/625 [==============================] - ETA: 0s - loss: 0.1431 - acc: 0.9504\n",
      "Epoch 2: val_acc did not improve from 0.87740\n",
      "625/625 [==============================] - 112s 179ms/step - loss: 0.1431 - acc: 0.9504 - val_loss: 0.3410 - val_acc: 0.8756\n",
      "Epoch 3/10\n",
      "625/625 [==============================] - ETA: 0s - loss: 0.1106 - acc: 0.9618\n",
      "Epoch 3: val_acc did not improve from 0.87740\n",
      "625/625 [==============================] - 112s 179ms/step - loss: 0.1106 - acc: 0.9618 - val_loss: 0.3778 - val_acc: 0.8708\n",
      "Epoch 4/10\n",
      "625/625 [==============================] - ETA: 0s - loss: 0.0800 - acc: 0.9747\n",
      "Epoch 4: val_acc did not improve from 0.87740\n",
      "625/625 [==============================] - 112s 179ms/step - loss: 0.0800 - acc: 0.9747 - val_loss: 0.4230 - val_acc: 0.8710\n",
      "Epoch 5/10\n",
      "625/625 [==============================] - ETA: 0s - loss: 0.0669 - acc: 0.9800\n",
      "Epoch 5: val_acc did not improve from 0.87740\n",
      "625/625 [==============================] - 112s 179ms/step - loss: 0.0669 - acc: 0.9800 - val_loss: 0.4831 - val_acc: 0.8674\n",
      "Epoch 5: early stopping\n"
     ]
    }
   ],
   "source": [
    "r1 = model4.fit(X_train, Y_train, epochs=10, validation_split=0.2, callbacks=[es, mc])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "f3301242",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "782/782 [==============================] - 40s 51ms/step - loss: 0.5054 - acc: 0.8556\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[0.5053685307502747, 0.8555999994277954]"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model4.evaluate(X_test, Y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ddfb9979",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "330403ec",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d39059a1",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "collapsed_sections": [],
   "name": "과제1.ipynb",
   "provenance": []
  },
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}